Systems and Methods for Allocating Inventory in a Fulfillment Network

1. A method, comprising: performing, via one or more computers: selecting items to be stored at a fulfillment center, wherein said selecting comprises: identifying orders fulfilled by a fulfillment center during a previous time period, wherein each of the identified orders specifies one or more of a plurality of items, and wherein at least one of the identified orders specifies a plurality of distinct items; determining, for each identified order, a probability that the identified order will occur in a subsequent time period; calculating, for each identified order, a cost savings associated with fulfilling the identified order from the fulfillment center instead of another fulfillment center; selecting a limited number of distinct items to be stored at the fulfillment center, wherein said selecting comprises selecting items among the plurality of items to be stored in the fulfillment center during at least a portion of the subsequent time period based, at least in part, upon the determined probability and cost savings of each identified order, wherein said selecting attempts to maximize the cost savings for the limited number of distinct items; and for each of the selected items, calculating a target inventory level, wherein calculating the target inventory level comprises: receiving a customer demand forecast corresponding to the selected item; determining a bind factor for the selected item; and calculating the target inventory level for the selected item to be stored in the fulfillment center based, at least in part, upon the customer demand forecast and the bind factor.

2. The method of claim 1 , wherein the cost savings include a shipping cost savings.

3. The method of claim 1 , wherein said selecting items among the plurality of items to be stored in the fulfillment center is performed independent of forecasted demand for the item.

4. The method of claim 1 , wherein determining the probability further comprises: grouping each identified order into one of a plurality of order groups, wherein each of the order groups specifies a different set of the plurality of items; and dividing a number of orders in a given order group by a total number of identified orders.

5. The method of claim 4 , wherein selecting the items includes iteratively selecting one of the plurality of items at a time until the limited number of selected items is reached.

6. The method of claim 4 , further comprising selecting a subsequent one of the plurality of items after having a selected a previous one of the plurality of items based, at least in part, upon an additional cost savings credit given to the subsequent item in response to the subsequent item being part of an order group that includes the previous item.

7. The method of claim 1 , wherein the customer demand forecast is a probability distribution.

8. The method of claim 1 , wherein the bind factor is a number between 0 and 1.

9. The method of claim 1 , wherein determining the bind factor for the selected item further comprises dividing a number of units of the selected item shipped by the fulfillment center to a region assigned to the fulfillment center during the previous time period by a total number of units of the item stored in the fulfillment center during the previous time period.

10. The method of claim 1 , wherein calculating the target inventory level further comprises multiplying the customer demand forecast by the bind factor.

11. A system, comprising: at least one processor; and a memory coupled to the at least one processor, wherein the memory stores program instructions, and wherein the program instructions are executable by the at least one processor to cause the system to select items to be stored at a fulfillment center, wherein to select items to be stored at a fulfillment center the program instructions are further executable to: identify orders fulfilled by a fulfillment center during a previous time period, wherein each of the identified orders specifies one or more of a plurality of items, and wherein at least one of the identified orders specifies a plurality of distinct items that are ordered together; determine, for each identified order, a probability that the identified order will occur in a subsequent time period; calculate, for each identified order, a cost savings associated with fulfilling the identified order from the fulfillment center instead of another fulfillment center; and select items among the plurality of items to be stored in the fulfillment center during at least a portion of the subsequent time period based, at least in part, upon the determined probability and cost savings of each identified order.

12. The system of claim 11 , wherein the cost savings include a shipping cost savings.

13. The system of claim 11 , wherein to determine the probability, the program instructions are further executable by the at least one processor to cause the system to: group each identified order into one of a plurality of order groups, wherein each of the order groups specifies a different set of the plurality of items; and divide a number of orders in a given order group by a total number of identified orders.

14. The system of claim 13 , wherein the program instructions are further executable by the at least one processor to cause the system to iteratively select one of the plurality of items at a time until a specified number of selected items is reached to attempt to maximize the cost savings for the specified number of items.

15. The system of claim 13 , wherein the program instructions are further executable by the at least one processor to cause the system to select a subsequent one of the plurality of items after having a selected a previous one of the plurality of items based, at least in part, upon an additional cost saving credit given to the subsequent item in response to the subsequent item being part of an order group that includes the previous item.

16. A non-transitory computer-readable storage medium having program instructions stored thereon that, upon execution by a computer system, cause the computer system to: calculate a target inventory level, wherein to calculate the target inventory level the program instructions further cause the computer system to: identify an item to be stored in a fulfillment center, wherein the fulfillment center is configured to fulfill an order for the item within a given geographic region; receive a customer demand forecast corresponding to the item; determine a bind factor for the item, wherein to determine the bind factor the program instructions cause the computer system to estimate a historical bind factor that when applied to a target inventory level adjusts the target inventory level for past demand that was not fulfilled from the fulfillment center for an assigned region; and calculate the target inventory level for the item based, at least in part, upon the customer demand forecast and the bind factor.

17. The non-transitory computer-readable storage medium of claim 16 , wherein the customer demand forecast is a probability distribution, and wherein the bind factor is a number between 0 and 1.

18. The non-transitory computer-readable storage medium of claim 16 , wherein the bind factor is determined based at least in part on historical order data.

19. The non-transitory computer-readable storage medium of claim 18 , wherein to determine the bind factor, the program instructions, upon execution by the computer system, further cause the computer system to: determine a number of units of the item that were shipped from the fulfillment center to a region assigned to the fulfillment center during a preceding time period; determine a number of units of the item that were stored in the fulfillment center during the preceding time period; and divide the number of shipped units by the number of stored units.

20. The non-transitory computer-readable storage medium of claim 16 , wherein to calculate the target inventory level, the program instructions, upon execution by the computer system, further cause the computer system to calculate a product of the customer demand forecast with the bind factor.